# The Context Bible — Scholar Citation Engine

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [John Tad](https://cerebralvalley.ai/u/JXT)
- **GitHub:** https://github.com/johtad/context-bible-scholar-agent
- **Demo video:** https://www.youtube.com/watch?v=Xoc3q7JtANo
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1131

Millions of people now ask AI their Bible questions — but a large share of serious Bible readers reject the answers on principle. They do not want a machine summarizing or interpreting Scripture itself; to them, that is ground a model should not occupy. Every current AI Bible tool crosses that line: chatbots and basic RAG apps paraphrase the text in their own voice, blend sources without attribution, make claims nobody can check — and when they have nothing solid to draw on, they answer anyway. For the more than 100,000 readers of The Context Bible, that makes today's AI Bible tools unusable — fluent, but unverifiable and out of bounds. What does not exist is an AI Bible tool whose every claim is attributed to a named, published scholar and machine-checked against that scholar's work before anyone sees it.

The Context Bible — Scholar Citation Engine answers Bible questions the way a careful research librarian would: it never speaks for the biblical text itself. Instead, an agent built on Muse Spark retrieves what named, published scholars have written about a passage — D. A. Carson, Raymond E. Brown, John Goldingay, Derek Kidner, and others — and drafts an answer only from those retrieved sources, with an inline citation attached to every claim. A second Muse Spark pass then fact-checks the draft claim by claim against the sources; anything unsupported is struck, the answer is regenerated, and the result re-verified before the user sees it. The interface shows the whole trail: the agent's tool calls, the citations, and the verification report.

Two things this is not. It is not basic RAG: there is no chunk store and no single stuffed prompt — the model works as an agent, identifying the passage and retrieving structured, attributed scholar entries through tool calls, then drafts under a hard rule that every claim must trace to a retrieved entry. And it is not an education chatbot: it never teaches or interprets in its own voice — it has no voice of its own, only the scholars'. Where a chatbot always answers, this system verifies its draft claim by claim and, when no scholarship is retrieved, declines to answer at all.

The design answers a real conviction in the audience I serve. I teach at Pepperdine University and built The Context Bible, a Bible-study app with a Facebook community of over 100,000 readers. Many of them will not use an AI that summarizes Scripture — they consider that ground a machine should not occupy. An agent that cites scholarship, and can prove its citations, earns trust a chatbot cannot. The engine currently covers John 14 and Psalm 91 from a curated library built during this hackathon; the format extends to any passage and any commentary library.

## More from Meta Global AI Developer Hackathon

- [Sunny Sonnendeck](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1128)
- [Orbit](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1129)
- [KRAKADYL](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1130)
- [ROOKWELD](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1132)
- [PuzzleAR](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1133)
- [Iqbal's Basilisk Lab](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1134)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1131. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt